Thursday, 1 September 2016

Jowett flat-four engines: Early development

During early prototype testing of the Javelin it became apparent that the engine was mechanically noisy in a harsh resonating manner. This was deemed to be a major problem given that the Javelin was to be a luxury small family car and as such a remedy was required. Two types of crankcase had been manufactured, one in cast-iron, the other in aluminium. Experimental testing on these early designs made it clear that there was considerable whirl of the crankshaft and a certain amount of flexing on the crankcase was evident from edge marking on the bearings.

The next stage of development was to manufacture a cast-iron crankcase and a three bearing crankshaft, the bearing-cap joint faces now being horizontal, so the crankshaft could be dropped out of the bottom of the engine. The first engine tested of this type was a 1200cc unit, but when experiments were carried out with larger diameter liners, increasing the capacity to 1500cc, considerable crankcase ‘thump’ was experienced. To overcome this for testing purposes a boiler plate was bolted across the bottom of the bearing-caps, however this only highlighted the inherent weakness of this crankcase design. Subsequent development led to the adoption of light-alloy crankcases split vertically, which permitted the use of tie-bolts, making the whole assembly much stiffer.
Alloy crankcase of the production Jowett flat-four

Jowett had noted that cast-iron crankcases resulted in a quieter engine, but it was decided that the alloy crankcase should be proceeded with as it had been designed for die-casting. The cast-iron version was approximately 10% quieter than the alloy one, but was naturally heavier. In view of this and the difficulty at that time of obtaining iron castings, the spilt alloy crankcase was decided upon. It was also at this stage that the 1200cc project was dropped, as there was a large performance difference between it and the 1500cc engine.

When the revised engine was power tested, a considerable drop in performance was seen above 4250rpm. This was attributed to inadequate breathing and poor turbulence in the stepped head. The valve lift was increased from 0.275in to 0.315in and the ports ‘cleaned up’. Weslake was called in to inspect the combustion chamber and he evolved a semi-pancake head with 14mm plugs easier to produce and increasing top-end power by 15%, while providing smoother running. The exhaust system was changed from streamlined exhaust ports brought out to the bottom face of the head, to a manifold bolted to the underside of the head, the off-side manifold feeding into a pipe running round the front of the engine to enter the near-side manifold and take benefit thereby of extractor action. The main exhaust pipe led from the back of the near-side manifold. It had 1 3/8in inside diameter and the power drop with silencers was only 3 BHP compared with an open pipe. This new exhaust arrangement gave a power increase of 1.5% and resulted in the cylinder heads no longer being handed, a production and servicing advantage.

Jowett engine sectioned

Snatchy running below 20mph lead to an increase of flywheel diameter to the limits of the bell-housing. Another alteration required following testing, was to change the main bearing clearance due to rapid crankcase expansion. A steel housing giving 0.0003in to 0.0018in clearance at assembly temperature was adopted.

The Javelin emerged as the first really new British post-war car. It was a comfortable, brisk 5/6 seater saloon giving 75/80 mph and 28/32 mpg with advanced aspects such as the flat four engine, torsion-bar suspension and wind-defeating body form

The prototype engines, developing 40-45 BHP had been satisfactory in respect of bearings, but long-distance driving on the Continent with the early production versions showed up a tendency to run big-end and main bearings.

With the previously mentioned improved breathing 50-52 BHP was developed at 4500rpm, and it was decided that white-metal bearings must be replaced by copper-lead bearings, if possible in conjunction with the existing EN 12 steel crankshaft. The flat-four layout led to higher oil temperatures than are experienced in in-line designs, which contributed to the bearing failures.

Initially sintered copper-lead bearings with a white metal flash of 0.00125in were utilised with the un-hardened steel crankshaft. These bearings showed no signs of fatigue, but were extremely sensitive to dirt and scuffing on the crankshaft. A hardened crankshaft was therefore adopted with special care being paid to assembly and running-in. It was also found that the stepped location on the big-end led to distortion on tightening, so a new con-rod was devised, the big-end having an offset serrated face and clamp bolts increased to 0.375in and 400lb/in tightening torque. A dirt trap hole of 1/16in diameter had a negligible effect on oil pressure and consumption. The crankshaft was induction hardened on the journals and pins to a hardness of 512-530 Brinell, and the bearing surfaces were lapped to a finish of 8-12 micro-inches against the former 12-24 micro-inches. A softer bearing material of 30% lead, 1.2% tin and 68.8% copper with a 0.00025 in plated white-metal layer for running-in was used with the new rods and crankshaft and the bearings now stood up to 50 BHP and 4750 rpm in spite of the higher oil temperatures and compact bearings of the flat-four layout.

The lubrication system was thoroughly tested in the initial stages of development, an engine being rigged for measurement of oil spillage from bearings, relief valve, ancillary services etc. As a result, the feed to the main bearings was increased and the size of the oilways increased to 7/16in diameter to obviate a possible danger of bearing starvation under cold-start conditions with the full-flow filter system adopted to ensure clean oil for the hydraulic tappets. The pressure relief valve exhausted beneath the sump oil level to avoid aeration and later the discharge was by-passed to the pump suction side, within the cover.

Initially the three-bearing crankshaft engine suffered with oil swirl due to air transferring from one side of the crankcase to the other. To prevent this, a surface baffle was fitted which allowed the free passage of air only. Originally the oil pump had been mounted on a bearing cap, but the vertically split cases obviated this location, so it was moved to the timing-case wall and driven by spiral gears from the crankshaft. The distributor was also repositioned to allow it to be mounted vertically and use a common drive-shaft as the oil pump from the spiral bevel gear on the crankshaft. The oil pump capacity was also increased which meant oil pressure rose from 50lb/ to 65lb/ Following use on Jowett’s competition cars, an oil-cooler built to Jowett’s specification was incorporated on production engines in 1952. Initially this was placed rewards for accessibility, but later was moved to a location between the fan and radiator. With the oil-cooler in circuit, pressure pulsations occurred at audible frequencies until the previously mentioned dirt-trap holes in the big-end caps were deleted.

In vehicle testing it was shown that louvers in the bonnet became ineffective in terms of extracting air from behind the radiator above 50mph. Pressure areas were checked and it was found possible to take air from behind the radiator via apertures in the front wheel arches.

An issue with early production engines was noisy valve gear, even with the zero lash hydraulic tappets. Improved manufacturing tolerances were introduced, but the noise was still deemed unacceptable. Jowett then went on to investigate the effect of valve opening/closing ramp sizes and also velocity profiles. After much experimentation Jowett settled on cams with a 0.008in opening ramp and 0.020in closing ramp (from the fierce initial 0.002in and 0.006in ramps on the inlet and exhaust respectively). Unfortunately hydraulic tappets became unobtainable during 1950 and the noise level rose somewhat with the enforced use of conventional tappets.

Experiments were made with the material for camshafts and tappets. Excellent results were found with a high duty 1% chromium cast iron camshaft with a tip hardness of 40-45 Rockwell C and chilled iron tappets of a similar hardness and a finish of 7-10 micro inches. A phosphate process on cam and tappet faces to retain oil during running-in was found to be beneficial, but not necessary.

Five different types of liner/piston combinations were used in the course of development. Vacrit high-duty manganese chromium iron liners with a 280-270 Brinell surface hardness were originally used, in conjunction with split-skirt pistons in LO-EX or LM13 alloys withtwo D/26 radial thickness pressure rings and a slotted oil-control ring.

A taper-faced Vacrome chromium-plated top piston ring was adopted to cut oil consumption, without complete success. Liner distortion was suspected and investigation showed that whilst 0.008-0.010in gasket nip at 38/40 lb/ft cylinder head tightening torque was satisfactory to retain gas and water seals, this was highly critical; any degree of higher torque loading or excessive nip caused local liner collapse and consequently distortion. To counteract this, the liner section was stiffened and an internally-stepped second ring fitted to facilitate quick bedding-in of the chromium-plated piston ring. Following this a Javelin ran 80,000 miles in the course of testing by Avon India Rubber Co. Ltd., gave an average of 3,700 mpg of oil at 37.39 mph average speed, and maximum bore wear averaged 0.002in, equal to 40,000 miles per thou.

Due to the Javelin’s unusual firing order of 1-3-2-4 carburation was paid special attention. Cylinder 1 and 3 are fed from one carburettor via siamesed ports, and 2 and 4 from the other carburettor. To prevent a weak mixture in the front two cylinders of each bank caused by inlet tract surge, a 0.55in diameter balance pipe was added between the two carburettors. Intake noise was a problem on the Javelin and Jowett went on to test many different types of air filter and silencer, but no satisfactory solution was found. Instead Jowett evolved their own baffle box which was located in the alligator-bonnet, tuned to length to suit the induction system, and connected to a resonance chamber which was coupled to the air intakes by vertical pipes having squash rubber connections which broke as the bonnet was lifted. A non-spill oil-bath air filter was incorporated.

In addition to the early bearing failures and excessive oil consumption, gasket blowing was an issue on some Javelins. It was subsequently found that this was due to a too small an asbestos content at the fold of the gasket, however it was only with the increased output for competition purposes that this cause was identified.